A comparative analysis of the results of the 2014 and 2019 Moscow City Duma elections reveals that the unexpected result of 2019 was triggered by Alexey Navalny's "Smart Voting" strategy. The author makes the case that now the party system in Moscow is mostly composed of two factions: city administration and its opposition. The paper gives an overview of apportionment patterns in the Moscow City Duma depending on assumed redistribution of votes that has occurred in the five years since the 2014 election.
This paper is a remake of a blogpost I published on September 16, 2019 shortly after the Moscow City Duma election . Since my post prompted a number of research papers concerning the "Moscow City Duma Election-2019: What Was That?" subject, I was offered to rewrite it into a journal article. The edits I have made include adding an important segment and visual aids as well as trying to scale down the political science aspect.
The latter was achieved only barely, as my entire post rested on the statement that the political framework in Moscow is best described as a two-party system – Party A and Party N – instead of a multi-party system (United Russia, CPRF, A Just Russia, LDPR, etc.). Party A is the city administration party (although a better term would be "party of the elite"), while Party N is a party of politically active citizens (although the "N" technically stands for "Navalny," most of the members of this party do not in fact support Alexey Navalny).
My findings are based on comparing the results of the 2014 and 2019 Moscow City Duma elections. The public segment of the State Automated System "Vybory" (Rus. for "elections") holds extensive statistical data on the two elections. In this research, we used party results data broken down by constituencies, as their delimitation was the same for both elections (naturally, the number of voters by constituency changed; the change is indicated in Table 1 below; we assume, however, that these changes are barely significant for the findings presented in this paper).
First, it is worth noting that Party A has been systematically winning the Moscow City Duma elections since 1997. In the 1997 and 2001 elections, it was informally dubbed "Luzhkov's list" (winning by 27:8) and "The List of Four" (33:2) respectively. In 2005 (28:7), 2009 (32:3) and 2014 (38:7) elections, Party A ran under the banner of United Russia.
The 2019 presented an unusual situation: administrative "self-nominees" won 25 seats out of 45. Could it be that over those five years the opposition – CPRF, A Just Russia and Yabloko in this case – have gained the voters' recognition they are so proud of now? Alas, comparing electoral statistics from 2014 and 2019 indicates that their success is a consequence of protest voting, which was rallied by "Smart Voting" and made worse by the lack of interest in the election among citizens.
There was a serious miscalculation on administrative political strategists' part. Over the last 17 years they have been tweaking both electoral legislation and electoral law enforcement to maximize the success of the administration party, transforming the election into self-replication of government.
The tweaking was especially obvious in Moscow. In 2014, Moscow authorities pushed the federal legislator to allow to make all 45 constituencies single-member. Single-member constituencies allowed Moscow administration to easily get rid of competition, since proportional representation in a city as political as Moscow would let unwelcome elements into the docile Moscow City Duma.
The protest has obviously spread since 2014. Moscow administration's (as well as its Kremlin supervisors') skill at political spinning remained almost the same, however. The 2019 election was to be conducted based on the same "negotiated" technology as in 2014, but with the difference that the tired party brand would be discarded. We presume that should the procedure have remained completely the same, communists would have won a couple more seats (in addition to those they were initially allotted) and a couple constituencies would have gone to Yabloko. All because the extra protest votes that emerged after opposition candidates had been illegally excluded would have once again scattered among the competitors of the poorly disguised administration candidates.
But "it all went wrong" when other political strategists interfered and outmaneuvered their colleagues from Moscow. It must be stressed, however, that this was not a political, but a technological victory, as the two-party structure described above only works in elections held when a confrontation between the state and the society is ongoing. It does nothing for either political or legislative working processes. Party N deputies (with few exceptions perhaps) are not likely to admit they owe their seat to "Smart Voting." There will be no smart voting in Moscow City Duma.
Let us compare the 2014 and 2019 electoral statistics. In both 2014 and 2019, the candidates were nominated in 45 single-member constituencies, with both parties and independents being allowed to be nominated. By speaking of party results, we mean the results of the candidate nominated by the party. One exception: in 2014, Party A candidates (meaning those backed by administration at the time of the election) were largely (in 29 cases out of 39) nominated by United Russia, while in 2019 United Russia formally nominated no candidates.
The fact that the candidate indeed belonged to Party A was determined based on the records published by the media outlet "Meduza"  and subsequently checked with the data on Dmitry Baranovsky's website (the site was later blocked by Roskomnadzor) . The list of "Smart Voting" (SV) candidates – meaning those recommended by Alexey Navalny – was obtained from The Bell news agency . Since certain records on whether or not the candidate belongs to Party A may be disputable, below is a list of candidates we used in our research (the number before the candidate's last name is the constituency number; there were no Party A candidates in the constituencies No. 20, 21, 37 and 43).
1. Titov; 2. Volovets; 3. Tsvetkova; 4. Kiselyova; 5. Babayan; 6. Belykhin; 7. Perfilova; 8. Kopeikina; 9. Medvedev; 10. Kartavtseva; 11. Nifantyev; 12. Shaposhnikov; 13. Buskin; 14. Pochinok; 15. Metelsky; 16. Molev; 17. Tatulova; 18. Tabashnikov; 19. Nazarova; 22. Svyatenko; 23. Nikolayeva; 24. Dyagilev; 25. Stebenkova; 26. Shchitov; 27. Orlov; 28. Samyshina; 29. Artemyev; 30. Rusetskaya; 31. Zverev; 32. Melnikova; 33. Guseva; 34. Semennikov; 35. Metlina; 36. Sharapova; 38. Kozlov; 39. Golovchenko; 40. Batysheva; 41. Gerasimov; 42. Nikitin; 44. Sviridov; 45. Kasamara.
- Party A – 25 seats;
- CPRF – 13 seats;
- A Just Russia (JR) – 3 seats;
- Yabloko – 3 seats;
- Independent – 1 seat (Daria Besedina, a member of Yabloko, running as an independent).
The SV list result – 20 seats.
All 20 winning candidates were on the "Smart Voting" (SV) list. Figure 1 shows winning candidates endorsed by SV classified by parties.
All winning candidates nominated by JR, Yabloko as well as an independent were endorsed by SV.
24 out of 25 runner-up candidates were on the SV list.
Naturally, a question begs to be answered: what is the effect and what is the cause of such an unusual result? Was the growing protest the main contributing factor to the opposition's victory while SV simply guessed the most promising candidate right? Or was it the opposite: SV's contribution was significant and the opposition's results would have been much worse without it?
To answer this question, let us take a look at how the results of party candidates changed compared to 2014.
It is quite obvious with CPRF, A Just Russia and LDPR, since their candidates were in the ballots in both 2014 and 2019. Making a similar comparison for Yabloko is challenging, as their candidates made it into the ballots of 44 constituencies in 2014, and only of 3 in 2019.
Let us take a look at the parties' achievements in all constituencies over the past five years.
Figure 2 shows an increase in votes that a CPRF candidate achieved constituencies-wise over the past 5 years (there were no communist candidates in constituencies No. 3, 14 and 45 in 2019). The constituencies are sorted in ascending order for illustrative purposes. Diagonal crosshatch markings define the constituencies where the CPRF candidate was endorsed by "Smart Voting."
"Smart Voting" endorsed CPRF in 33 constituencies. Over five years, the increase in CPRF candidate's result in all these constituencies ranges between 5.7% and 29.4%. Without SV, the result decreased in six constituencies (by a max of 19.2% in the 26th constituency) and increased only in three constituencies (by a max of 12.3% in the 13th constituency).
Figure 3 shows the increase for candidates from the four parties in all 45 constituencies over the past five years. Crosshatched are the increases for the parties with SV-endorsed candidates (there is no bar on the graph for the party in the constituency if the party's candidate was not in the ballot in either 2014 or 2019).
A Just Russia was endorsed by "Smart Voting" in 8 constituencies. JR’s increase in all these constituencies ranged between 5.4% and 29.4%. Without SV, the result increased slightly (by a max of 8.5%) in 15 constituencies and decreased (by a max of 18.4% in the 6th constituency) in 15 constituencies as well.
Comparing Yabloko's advances is more challenging, as in 2019 Yabloko nominated candidates in 15 constituencies only, with 12 candidates denied registration. Note that in 2014, Yabloko candidates were on the ballots of 44 constituencies and gained between 3.4% and 27.9% votes, without winning any seats. This time, all three registered Yabloko candidates were endorsed by SV and became deputies. That said, Yevgeny Bunimovich improved Sergei Grigorov's result by 16.6%, Maksim Kruglov improved his own result by 23.3% while Sergei Mitrokhin improved Sergei Ivanenko's result by 33.4%.
As for Daria Besedina, the SV-endorsed independent candidate in the 8th constituency, she managed to outperform all party members, drop A Just Russia's increase to 2.8% and drive the increase of CPRF and LDPR into negative.
While looking at Figure 3, a few things are especially obvious:
- in all constituencies, SV improved (compared to 2014) the result of the party it endorsed;
- in all constituencies, the increase in votes for SV-endorsed parties is much stronger than that of other parties (including United Russia). These increases can only be compared in constituencies No. 13, 28 and 35, where the SV-endorsed candidate from A Just Russia became the runner-up.
Moreover, the increase for other election participants (by force of independents and nominees from the party "Communists of Russia") is higher than that of SV-endorsed candidates in 3 constituencies only (no. 24, 30 and 35). That said, independent candidates are largely responsible for the increase: Andrei Tarasov for 2/3 in constituency no. 24, Roman Yuneman for more than 3/4 in constituency no. 40 and Sergei Malakhov for nearly half in constituency no. 35.
There is a recurrent hypothesis that the increase for "opposition" parties (meaning all parties except Party A) stems from the growing protest, while all "Smart Voting" did was guess for whom it is going to be highest. In this case, we have to admit that SV authors have excellent forecasting skills: they even managed to predict the growth pattern of protest in at least 42 constituencies out of 45.
It is anyone's guess what would have happened if the voters had not heeded the appeal of "Smart Voting." It is natural to assume, however, that the additional protest votes would have distributed more evenly than the "Smart Voting" ones. It is possible to calculate what would have happened given the more even distribution of additional votes, although a distribution model would have to be provided first.
Candidates nominated by CPRF, LDPR, JR and Yabloko (we will refer to them as "opposition" parties) could have received additional protest votes from the following sources:
- additional turnout increase (if any);
- votes overflown from United Russia;
- votes overflown from other parties and independent candidates;
- votes overflown from other opposition parties.
Each of these sources could make a negative contribution, meaning the votes would actually come to the source instead of leaving it (turnout growth could be negative, like in constituency no. 17, for example; the number of votes for Party A could increase, like in constituency no. 2, for example).
Calculating the increase (or the decrease) in votes in each constituency and based on each source for the 2019 election is easier than for the 2014. Below is the table resulting from such calculations (where every row sums to zero).
|Constituency||Decrease in voters (increase with the sign reversed)||Party A||CPRF||LDPR||JR||YABLOKO||Others (including the excluded)|
The table is not fully displayed Show table
Candidates nominated by certain parties may not have been in the ballots in either 2014 or 2019, or in neither of these years (this, in fact, explains the large negative numbers in Yabloko column as well as the zeroes in the table).
In every constituency, the votes that could have otherwise gone to CPRF, LDPR, JR and Yabloko if their candidates were in the ballots are composed of the votes that came from the sources described above. That said, the increase might turn out to be negative. This number can be calculated for each constituency. Table 2 shows the number of votes that could have been added to the opposition candidates represented in the ballot by force of turnout increase, other parties and independent candidates (the calculation was made by adding up all negatives in the row that correspond with the sources whose candidates were not in the 2019 ballots).
|Constituency||Parties with candidates in 2019 ballots||Additional votes total|
|1||CPRF, LDPR, JR||5873|
|2||CPRF, LDPR, JR||9087|
|4||CPRF, LDPR, JR||8185|
|5||CPRF, LDPR, JR||8091|
|6||CPRF, LDPR, YABLOKO||10156|
|7||CPRF, LDPR, JR||8659|
|8||CPRF, LDPR, JR||1438|
|9||CPRF, LDPR, JR||9555|
|10||CPRF, LDPR, JR||6993|
|11||CPRF, LDPR, JR||1112|
|12||CPRF, LDPR, JR||8583|
|13||CPRF, LDPR, JR||9831|
|14||LDPR, JR, YABLOKO||12775|
|16||CPRF, LDPR, JR||7592|
|17||CPRF, LDPR, JR||11189|
|18||CPRF, LDPR, JR||8834|
The table is not fully displayed Show table
Note the unexpected decrease in protest votes in constituencies no. 29 and 38. This phenomenon requires further study.
To understand what would have happened had the votes distributed in a different pattern, we shall study two distribution models for each constituency.
Model 1: additional votes are distributed between the registered candidates from "opposition" parties in proportion to the share of votes the "opposition" party (represented by the nominated candidate) received in the 2014 election among the same parties.
Model 2 is more extreme: all additional votes go to the candidate from the opposition party that gained the most votes in the given constituency in 2014.
Model 2 is barely realistic: it virtually means that all "opposition" parties are consolidated. However, it does demonstrate the impact of "Smart Voting" on the result of the election as much as Model 1.
The assumptions of both models provide different election results in the constituencies (Table 3) and different seat allocation in Moscow City Duma of the seventh convocation (Table 4 and Figure 4).
|Constituency||The party (or the independent candidate) that won a seat under Model 1||The party (or the independent candidate) that won a seat under Model 2||Seat allocation in Moscow City Duma of the seventh convocation (asterisk indicates the SV-endorsed parties)|
The table is not fully displayed Show table
The additional winning independents in the assumptions made in the models stem from the fact that in the 2014 election, independent candidates received a major share of votes in their constituencies.
Although the assumptions are subject to criticism, the models still speak volumes. They validate the finding that "Smart Voting" indeed helped the "opposition" parties to gain additional seats. Compared to the more realistic Model 1, "Smart Voting" gave 9 more seats to CPRF and 2 more seats to JR and Yabloko each. Even compared to the unrealistic Model 2, which implies consolidation of all opposition parties, "Smart Voting" gives 1 more seat to JR and 3 more seats to Yabloko.
Note that it is possible to model other blocs of "opposition" parties, although they all will indicate that SV favors any "opposition" party.
The given analysis points to a conclusion of varying degrees of extreme:
A) "Smart Voting" enabled CPRF, JR and Yabloko to appoint extra deputies to the Moscow City Duma;
B) "Smart Voting" significantly changed the composition of the Moscow City Duma;
C) "Smart Voting" won this election;
D) this election was a victory for Alexey Navalny (who does not have passive suffrage) and his team.
In support of the last statement I shall provide the results of the 2013 Moscow mayoral election (without including invalid ballots): back then, Party A (meaning Sergei Sobyanin) received 52.2 votes, Navalny received 27.7%, and all opposition parties received 20.2% combined. Altogether, the opposition parties and Navalny received 47.8% votes. Now is the time for a heads-up.
In the 2019 election, Party A deputies received 52.9% votes combined, while the winning candidates received 47.1% (to calculate this data, the number of votes received by all winning Party A candidates should be divided by the number of votes received by all winning candidates; the second number is calculated by subtracting the first number from 100%). As was assumed, a stable electoral configuration has formed in Moscow over the past 5 years, although it does not consist of United Russia, CPRF and others, but is instead formed by two unregistered parties: Party A and Party N.
"Smart Voting" only worked because the majority of Russian citizens do not take any interest in the election. 78% did not go to the polls. This is the result of what has been happening to elections in Russia over the past 20 years or so. Party A tried its hardest by offering gifts at polling stations and putting all clients of social security departments on the voting-at-home electoral rolls. The higher the level of voter apathy, the higher the level of voting at home in Russia. Below is the statistics of voting outside the polling place (the percentage of the overall turnout):
|Voting at home percentage||2.8%||4.6%||4.1%||5.2%||5.8%||7.4%|
The turnout was 22% after all. Many concerned people came to the polling stations as well as those who realized what "Smart Voting" was. The rest came in extremely small numbers. The result is obvious.
Another important thing is that the turnout was not inflated by 10 as it was in 2007–2011 in Moscow. The 2011–2012 protests scared the practice away, at least for now. Now the attention is on practicing to disperse the protests. Losing the political strategy round, however, might bring the old tricks back.
The author would like to thank Arkady Lyubarev for his valuable feedback and especially for pointing out the errors in the initial text.
Received 19.04.2020, revision received 06.05.2020.